WO2010050089A1

WO2010050089A1 - Encoding method and encoding device for compression encoding of moving images

Info

Publication number: WO2010050089A1
Application number: PCT/JP2009/002867
Authority: WO
Inventors: 田中康晴
Original assignee: パナソニック株式会社
Priority date: 2008-10-29
Filing date: 2009-06-23
Publication date: 2010-05-06
Also published as: JPWO2010050089A1; US8705628B2; US20110176027A1; CN102187678A

Abstract

A stop block detection unit (100) is provided in the front stage of a moving image encoding unit (200). The stop block detection unit (100) divides a frame to be processed into multiple blocks and sorts the blocks in the frame in order of increasing movement. The number of blocks designated by "relevant block setting number" input from the outside to an encoding device (300) are set as "stop blocks," starting with the one with the least movement. With respect to the blocks that correspond to stop blocks, the moving image encoding unit (200) executes inter-encoding for the moving image by forcibly setting "movement vector = 0" without performing a movement search, and forcibly substituting "DCT coefficient = 0" without performing DCT and quantification processing.

Description

Encoding processing method and encoding processing apparatus in moving image compression encoding

The present invention relates to a device having a moving image shooting function such as a digital camera or a camera-equipped mobile phone, and a moving image compression encoding technique when creating or using image content.

In recent years, moving picture compression coding technology such as MPEG (moving picture expert groups) that performs high-efficiency coding has been rapidly put into practical use, and is widely used in video cameras and mobile phones.

In encoding techniques such as MPEG, various encoding modes are defined in the standard. For example, in MPEG-4, an “intra-encoding mode” in which only an image within a screen of a frame to be encoded (hereinafter referred to as the target image) is used for encoding, and a frame that has already been encoded (hereinafter referred to as “encoding mode”) , An image region having a strong correlation with the target image is detected from the reference frame (hereinafter referred to as motion search), and only a difference value between the image after the motion search (hereinafter referred to as motion compensated image) and the target image is encoded. There is an “inter-coding mode”.

For the latter inter coding mode, DCT (discrete cosine transform) processing and quantization processing were performed on the motion vector representing the position of the reference image in the reference frame and the difference image between the input image and the motion compensated image. A method of generating a stream by variable-length coding a combination with later data DCT coefficients is shown in the MPEG-4 standard.

The inter coding mode will be described in more detail. When the input image and the motion compensated image match and the difference image is 0, 0 is input to the DCT process and the quantization process, and the result is output as 0. become. If there is no large difference even if the difference image is not 0, the result may be output as 0 by quantization. As described above, when the output result is 0, the DCT process and the quantization process are skipped without being executed, and the result is replaced with 0 to be output (see Patent Document 1). ).

JP 2004-328799 A

However, in order to determine whether or not the difference image is 0, that is, whether or not the DCT process and the quantization process can be skipped, a motion search is actually performed on each macroblock to obtain the difference image. Therefore, the number of macroblocks that can skip DCT processing and quantization processing cannot be grasped before frame processing is started. For this reason, there has been a problem that the degree of processing speed cannot be measured before the start of frame processing. For example, skipping DCT processing and quantization processing may be selected for all macroblocks in one frame, which may lead to higher speed, but DCT processing and quantization for all macroblocks in one frame. In some cases, skipping of the process was not selected at all and the speed was not improved at all.

An object of the present invention is to achieve a certain speed increase in frame processing without fail, thereby achieving downsizing and power saving of the encoding processing system.

In order to achieve the above object, a first encoding processing method according to the present invention divides a processing target frame into a plurality of blocks, and for each block, an input image and a reference image that has already been encoded are used. In an encoding processing method in which motion search is performed to generate a motion vector and a moving image encoding process is performed using the motion vector, S blocks (S is an integer equal to or greater than 1) out of blocks in the frame. A step of selecting a block, and a step of performing a video encoding process by forcibly replacing the motion vector with 0 without executing the motion search for the selected S blocks It is characterized by.

This makes it possible to create a block that does not necessarily execute a certain number of motion search processes in the encoded frame, and to realize high-speed processing.

The second encoding processing method according to the present invention divides a processing target frame into a plurality of blocks, executes DCT processing and quantization processing for each block, generates DCT coefficients, and converts the DCT coefficients into In the encoding processing method using the moving image encoding process, a step of selecting S (S is an integer of 1 or more) blocks from the blocks in the frame, and for the selected S blocks The DCT processing and the quantization processing are not executed, and the DCT coefficient is forcibly replaced with 0 and the moving image encoding processing is performed.

This makes it possible to create a block that does not necessarily execute a certain number or more of DCT processing and quantization processing in the encoded frame, thereby realizing high-speed processing.

The third encoding processing method according to the present invention divides a processing target frame into a plurality of blocks, and performs motion search for each block from an input image and a reference image that has already been encoded. In a coding processing method for generating a vector and performing a moving picture coding process using the motion vector, a first step of detecting a moving amount of an image and storing it in a moving amount storage unit, and the moving amount storage unit A second step of comparing the movement amounts of the blocks and selecting S (S is an integer of 1 or more) blocks from the smaller movement amount, and for the selected S blocks, the motion search And a third step of forcibly replacing the motion vector with 0 and performing a moving image encoding process, and performing the third step of a certain frame and simultaneously performing the process of the third step of the next frame Used in The amount of movement is characterized by obtaining in the first and second step that.

This makes it possible to further increase the processing speed by simultaneously executing the movement amount detection during the previous frame processing.

The fourth encoding processing method according to the present invention divides a processing target frame into a plurality of blocks, performs DCT processing and quantization processing for each block, generates DCT coefficients, and converts the DCT coefficients into In the encoding processing method using the moving image encoding process, the first step of detecting the moving amount of the image and storing it in the moving amount storage unit is compared with the moving amount in the moving amount storage unit, The second step of selecting S (S is an integer greater than or equal to 1) blocks from the least moving amount, and the DCT processing and the quantization processing are not performed on the selected S blocks. And a third step of forcibly replacing the DCT coefficient with 0 and performing a moving image encoding process, and performing the third step of a certain frame and simultaneously using the movement of the third step of the next frame Before quantity It is characterized by determining at the first and second step.

The first encoding processing apparatus according to the present invention divides a processing target frame into a plurality of blocks, performs a motion search from an input image and a reference image that has already been encoded for each block, and performs motion search. In an encoding processing apparatus that generates a vector and performs a moving image encoding process using the motion vector, the input image to be encoded is reduced to 1 / n (n is an integer of 1 or more). An input image reduction unit that generates an image, a reference image reduction unit that generates an image reduced to 1 / n with respect to a reference image that has already been encoded, the input image reduction unit, and the reference image reduction unit A movement amount detection unit that inputs a reduced input image and a reference image and detects a movement amount for each block, a movement amount storage unit that stores the movement amount detected by the movement amount detection unit, and the movement amount storage unit Movements remembered in The stop block selection unit that selects S blocks (S is an integer equal to or greater than 1) from the least moving amount, and the motion search is performed on the selected S blocks. And a moving picture coding unit that forcibly replaces the motion vector with 0 and executes a moving picture coding process.

The second encoding processing apparatus according to the present invention divides a processing target frame into a plurality of blocks, performs DCT processing and quantization processing for each block, generates DCT coefficients, and converts the DCT coefficients into An input image reduction unit that generates an image reduced to 1 / n (n is an integer equal to or greater than 1) with respect to an input image to be encoded; A reference image reduction unit that generates an image reduced to 1 / n with respect to a reference image that has already been encoded, and an input image and a reference image that are reduced from the input image reduction unit and the reference image reduction unit The movement amount detection unit that detects the movement amount for each block, the movement amount storage unit that stores the movement amount detected by the movement amount detection unit, and the movement amount stored in the movement amount storage unit are compared. From the side with the least amount of movement Stop block selection unit that selects blocks (S is an integer equal to or greater than 1), and for the selected S blocks, the DCT coefficient and the quantization process are not executed, and the DCT coefficient is forced And a moving image encoding unit that executes a moving image encoding process by replacing it with 0.

According to the present invention, a block with a small amount of movement is forcibly encoded as “motion vector = 0” and “difference image = 0”, and a certain number or more of the corresponding blocks are inserted in the frame. Therefore, since it is not necessary to perform the motion search process, the DCT process, and the quantization process for the corresponding block, the moving picture encoding process can be completed at high speed. In addition, a low power consumption effect can be expected by completing the processing at a higher speed.

1 is a configuration block diagram of an encoding processing apparatus according to Embodiment 1 of the present invention. It is a block diagram which shows the detailed structure of the stop block detection part in FIG. It is a figure which shows the 1st correspondence of the input image and reduced image in the encoding processing apparatus of FIG. It is a figure which shows the 2nd correspondence of the input image and reduced image in the encoding processing apparatus of FIG. It is a flowchart figure which shows operation | movement of the stop block selection part in FIG. It is a block diagram which shows the detailed structure of the moving image encoding part in FIG.1 and FIG.2. According to Embodiment 1 of the present invention, it is a diagram showing that a block at the same position may continue to be designated as a stop block for several frames. According to Embodiment 1 of the present invention, although the situation shown in FIG. 7 can be avoided by sequentially switching between a frame for which the number of corresponding block settings is set to 0 and a frame for which it is set to non-zero, the generated code amount It is a figure which shows that it becomes a result which fluctuates greatly. FIG. 9 is a diagram showing that stable bit rate control can be realized as compared with the situation of FIG. 8 by the second embodiment of the present invention for correcting the estimated code amount. FIG. 10 is a time chart showing that the stop block detection process and the moving picture encoding process are sequentially executed according to the first embodiment of the present invention, although the processing time is shortened as compared with the prior art. It is a structure block diagram of the encoding processing apparatus which concerns on Embodiment 3 of this invention. It is a block diagram which shows the detailed structure of the moving image encoding part in FIG. It is a time chart figure which shows that processing time is shortened according to Embodiment 3 of this invention compared with the case of Embodiment 1 by adoption of parallel processing. 1 is a block diagram illustrating a configuration of an imaging system using an encoding processing apparatus according to the present invention.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description of each embodiment, components having the same functions as those described once are given the same reference numerals and description thereof is omitted.

Embodiment 1
FIG. 1 is a block diagram showing a configuration of an encoding processing apparatus according to Embodiment 1 of the present invention. As shown in FIG. 1, the encoding processing device 300 includes an input image reduction unit 110, a reference image reduction unit 111, a stop block detection unit 100, a moving image encoding unit 200, and a reference frame storage unit 213.

<Function of each element>
The input image reduction unit 110 inputs an input image to be encoded, reduces the frame image to 1 / n (n is an integer of 1 or more), and outputs the frame image to the stop block detection unit 100. Similarly, the reference image reduction unit 111 also inputs a reference frame from the reference frame storage unit 213 in which a frame (reference frame) for which encoding processing has already been completed is stored, reduces the frame image to 1 / n, and stops. It outputs to the block detection part 100. In the stop block detection unit 100, a block composed of m × m pixels (m is an integer of 1 or more) using the input image input from the input image reduction unit 110 and the reference image input from the reference image reduction unit 111. The amount of movement of the image is detected in units. This movement amount detection will be described later.

When the movement amount detection of the image is completed, the blocks in the frame are sorted in ascending order of the movement amount, and are designated by the “number of corresponding block settings” input from the outside of the encoding processing apparatus 300 from the one with the smaller movement amount. Set the number of blocks as “stop blocks”.

The moving image encoding unit 200 receives stop block information input from the stop block detecting unit 100, the input image, and a reference image input from the reference frame storage unit 213, and performs coding according to the moving image encoding standard. Execute the conversion process. When executing the encoding process (inter-encoding process), the block corresponding to the stop block is forcibly set to “motion vector = 0” without executing the motion search, and the DCT process and the quantization process. Forcibly replacing “DCT coefficient = 0” without executing the video encoding process.

Thereby, in the block corresponding to the stop block, the motion search process, the DCT process and the quantization process can be skipped, and the process can be speeded up.

<Stop block detection>
FIG. 2 shows a detailed configuration of the stop block detector 100 in FIG. According to FIG. 2, the stop block detection unit 100 includes a movement amount detection unit 101, a movement amount storage unit 102, and a stop block selection unit 103.

The movement amount detection of the image performed by the stop block detection unit 100 will be described with reference to FIG. The input image is reduced to 1 / n by the input image reduction unit 110. The reference image is also reduced to 1 / n by the reference image reduction unit 111. This is input to the movement amount detection unit 101 in units of blocks of p × p (p is an integer of 1 or more) pixels.

In the movement amount detection unit 101, the input image of p × p pixels and the reference image of p × p pixels are input.
Movement amount = ΣΣ | input image (i, j) −reference image (i, j) | (1)
The sum of differences in pixel units expressed by Here, i and j are each an integer from 1 to p.

The above movement amount is calculated for each p × p pixel, and the result is stored in the movement amount storage unit 102. When the movement amount detection for one frame is completed, movement amount information for one frame is stored in the movement amount storage unit 102. The movement amounts are sorted in ascending order, and the sorted movement amounts are sent to the stop block selection unit 103 in ascending order. The stop block selection unit 103 uses the movement amount information sent from the movement amount storage unit 102 and the corresponding block setting number input from the outside, and selects which block (mx × m pixel) corresponds to the stop block.

Regarding the determination of the stop block, it is necessary to convert information calculated in units of p × p pixels with respect to an image obtained by reducing the frame by 1 / n into units of m × m pixels with respect to the same-size image of the frame. . In order to facilitate this, “p = m / n (the block size p × p of the reduced image is a value obtained by dividing the block size m × m of the equal-size image by 1 / n)” or “p = M (same processing unit for reduced image, same size image) ”, the conversion can be simplified.

In the case of “p = m / n” in the former case, one block that performs processing using the reduced image can be directly converted into one block of the same-size image on a one-to-one basis. Here, an example of “p = 8, m = 16, n = 2” is shown in FIG.

In the latter case of “p = m”, one block to be processed using the reduced image may be converted into an area of n × n blocks of the same size image. Here, an example of “p = 16, m = 16, n = 2” is shown in FIG.

As an advantage of the former, since the stop block determination is performed with the same number of blocks as that of the input image even in the reduced image, the determination can be performed in the minimum block unit. As the latter advantage, since the number of blocks of the reduced image can be reduced with respect to the input image, the memory capacity of the movement amount storage unit 102 that needs to store the movement amount information of all the blocks in the reduced image is reduced. can do.

The stop block selection unit 103 determines which block of the input image corresponds to one block of the reduced image output from the movement amount storage unit 102 using the relationship of p, m, and n described above, and stops the stop block. Control is performed so that the number of blocks of the input image corresponding to is at least equal to the number of blocks set. A specific processing flow will be described with reference to FIG.

First, all blocks in one frame of the input image are set as “moving blocks”. This moving block indicates that it is not a stop block, and in the block corresponding to the moving block, motion search, DCT processing, and quantization processing, which are normal encoding processes, are executed (S100).

Next, the movement amount information stored in the movement amount storage unit 102 is acquired one by one from the one with the smaller movement amount (S101). Based on the information, it is determined which input image block corresponds to the relationship of p, m, and n shown in FIGS. 3 and 4 (S102). A stop block is set (S103). For the block set as the stop block, motion search, DCT processing, and quantization processing are skipped in the subsequent encoding processing.

Finally, the total number of stopped blocks set in S103 is counted, and it is checked whether the value exceeds the number of corresponding blocks set (S104). When the total number of stop blocks exceeds the set number of corresponding blocks, the stop block determination process is terminated. If not, the process returns to S101 and a stop block is added.

In addition, although two kinds of relations were explained per the relation of p, m, and n in the above, it was taken as an example and does not necessarily need to be in that relation.

Further, in this example, the movement amount storage unit 102 has been described to store movement amount information of all blocks in one frame, but is necessary when the stop block selection unit 103 selects a stop block. Only information may be stored.

To explain with a specific example, when the stop block selection unit 103 reads the movement amount information of two blocks from the movement amount storage unit 102 according to the number of corresponding block settings, only the movement amount information for two blocks is stored in the movement amount storage unit 102. Can be stored. For example, a total of four blocks of movement amounts are written in the movement amount storage unit 102 in the order of “100”, “10”, “20”, “5” from the movement amount detection unit 101, and information “5” of two blocks among them is written. , “10” is transferred to the stop block selection unit 103, the movement amount storage unit 102 only needs to store information of two blocks having “5” and “10” as movement amounts. In this case, when data is written from the movement amount detection unit 101 to the movement amount storage unit 102, “100” is first input and “100” is stored. Next, “10” is input and has “10” and “100”. Next, “20” is input, “100” is erased, and “10” and “20” less than “100” are included. Finally, “5” is input, “20” is erased, and “5” and “10” are held. In this way, the capacity of the movement amount storage unit 102 can be made smaller by storing the blocks with a small movement amount so as to have only two blocks.

<Processing of blocks corresponding to stop blocks>
A moving image encoding process executed in response to stop block information output from the stop block selection unit 103 and an input image to be encoded will be described with reference to FIG. Since the present invention is effective in inter coding processing, an example of inter coding will be mainly described.

In a normal inter coding process, an input image is first input to the motion search unit 201. A frame image (reference image) that has already been processed is also input there, and a motion search process is executed using the input image and the reference image. The motion search result is passed to the motion compensated image generation unit 202, and a motion compensated image is generated from the reference image. Thereafter, a difference image between the input image and the motion compensated image is generated using the subtracter 203 and passed to the selector 204. The selector 204 selects an input image at the time of intra coding, and selects a difference image output from the subtracter 203 at the time of inter coding. In this example, a difference image is selected by the selector 204 to indicate inter coding. The difference image is input to the DCT processing unit 205 and subjected to DCT processing, and then the quantization processing is performed to the quantization unit 206 to generate DCT coefficients. The DCT coefficient is passed to the variable length coding unit 209 and variable length coding is executed. The motion vector information is also passed from the motion search unit 201 to the variable length encoding unit 209, and the motion vector information is also encoded. The variable length coding unit 209 generates variable length coding information including both motion vector information and DCT coefficients. The variable length coding information generated by the variable length coding unit 209 is output to the outside as a generated stream.

The DCT coefficient generated by the quantization processing unit 206 is also passed to the inverse quantization processing unit 207 and subjected to the inverse quantization process, and then passed to the inverse DCT processing unit 208 to perform the inverse DCT process. . The adder 212 adds the resultant data and the motion compensated image supplied from the motion compensated image generation unit 202 via the switch 211 to generate a reference image, and stores the reference image in the reference frame storage unit 213. .

The processing in the block determined to be a stop block with respect to the above normal inter coding processing will be described below. The inter coding process is executed in units of blocks, and whether the block to be executed is a moving block or a stop block is notified to the moving image encoding unit 200 as stop block information in units of blocks.

For the block notified as a stop block, the motion search process performed by the motion search unit 201 is not executed, and the execution result (motion vector) is set to 0 and output to the motion compensated image generation unit 202 and the variable length encoding unit 209. . Also, the DCT processing performed by the DCT processing unit 205 and the quantization processing performed by the quantization processing unit 206 are not performed, and the quantization processing result (DCT coefficient) is set to 0, and the variable length coding unit 209 and the inverse quantization unit It outputs to 207. Naturally, neither the inverse quantization unit 207 nor the inverse DCT processing unit 208 performs the process, and the inverse DCT processing unit 208 outputs 0 to the adder 212. In such a block set as a stop block, the execution of motion search processing, DCT processing, quantization processing, inverse quantization processing, and inverse DCT processing can be skipped.

As shown above, with respect to the blocks in each frame, a certain number of blocks or more are intentionally set as stop blocks, and the processing is always skipped for the stop blocks. Compared with performing normal encoding processing on the blocks, it is possible to achieve a certain speed increase. By using this function, high resolution processing or high frame rate processing that requires high-speed processing can be realized with a smaller mounting scale. In addition, since the process is executed at high speed, the process can be completed quickly, leading to low power consumption.

In the present embodiment, square blocks such as px × p or m × m are used, but rectangular blocks such as px × py and mx × my (px, py, mx, my are integers of 1 or more) may be used. . Also, for n representing a reduction ratio, it is possible to prepare reduction ratios by direction such as horizontal nx and vertical ny (nx and ny are integers of 1 or more) and perform reduction processing according to the prepared reduction ratios.

In the present embodiment, regarding the block set as the stop block, both the motion vector and the DCT coefficient are set to 0. However, only one (the motion vector only is 0 or the DCT coefficient is only 0). As a matter of course, it is possible to achieve sufficiently high speed even by a method in which only the motion search process is not executed, or the DCT process, the quantization process, the inverse quantization process, and the inverse DCT process are not executed.

<< Embodiment 2 >>
In the second embodiment, an implementation content that is a further development of the content explained in the first embodiment will be described.

In the first embodiment, a certain number of stop blocks are created for each frame to achieve a certain high speed and low power consumption. However, as shown in the example of FIG. Is continuously specified as a stop block for several frames, it is forcibly encoded as “motion vector = 0” and “difference image = 0”, so the same image as the previous frame is always displayed over several frames. It will continue to output. There is no problem if the image is stopped without moving even one pixel, but the amount of movement calculated by Equation 1 is smaller than the other blocks even though the image is moving several pixels at a time. When a stop block is designated, the fact that an image is updated only once every few frames (not designated as a moving block) means that it may lead to degradation of image quality as a moving image. In the second embodiment, implementation contents for solving this problem will be described.

As described in the first embodiment, the corresponding block setting number that defines the number of blocks to be a stop block in each frame is input to the stop block detection unit 100. The above problem is solved by switching the value of the number of corresponding block settings for each frame.

In a certain frame (qth frame), the number of corresponding block settings is set to non-zero so as to obtain a speed-up effect. In the next frame (q + 1 frame), the corresponding block setting number is set to 0, and all blocks are encoded as moving blocks. In the next frame (q + 2 frame), the corresponding block setting number is set to non-zero, and the speed-up effect is obtained as in the q frame.

In this way, by switching the frame setting number of the corresponding block to 0 and the frame setting to non-zero in units of frames, it is possible to obtain a speed-up effect by always embedding a fixed number of stop blocks. In addition, it is possible to prevent the same block described with reference to FIG. 7 from being continuously designated as a stop block over several frames.

In the above description, the example of providing a frame that sets the corresponding block setting number to 0 once every two frames has been described as an example. However, the corresponding block setting number is set to 0 once every r frames (r is an integer of 2 or more). Of course, other frames can be non-zero.

Also, as described above, when the number of corresponding block settings is switched in frame units and the encoding process is executed, the code amount of the frame greatly increases and decreases as shown in FIG. Of course, the code amount increases when the corresponding block setting number is 0, and the code amount decreases when it is non-zero. Thus, for example, when the code amount of the previous frame and the quantization parameter of the previous frame are used for bit rate control, the following problem occurs.

Generally, when the code amount is small in the previous frame, the code amount is controlled to increase in the next frame. On the other hand, when the code amount is large in the previous frame, control is performed so that the code amount is decreased in the next frame. However, if the increase / decrease is repeated as shown in FIG. 8, stable bit rate control cannot be performed. In other words, even if the same quantization parameter is set, there are frames with a large code amount and frames with a small amount of code, making it difficult to control the bit rate.

この This problem is dealt with by the method shown in FIG. In the example of FIG. 9, “the total number of blocks in a frame = 100” is set, and a frame in which the corresponding block setting number is 0 and a frame in which a half block (= 50) in the frame is switched are encoded every frame. Yes. In the latter frame in which “the number of corresponding block settings = 50”, only about ½ of the blocks are encoded, so the code amount of the frame is also halved. On the other hand, the estimated code amount when all the blocks in the frame are encoded as moving blocks is “code amount × total number of blocks in frame / number of corresponding blocks set (= code amount × 100/50 = code amount × 2). ". In the example of FIG. 9, the actual code amount is multiplied by a coefficient “2”. By using the estimated code amount value obtained in this way for bit rate control, an apparently equivalent code amount is output for the same quantization parameter, so more stable bit rate control Can be realized.

The above is provisionally stipulating the code amount of the previous frame used when determining the quantization parameter of the next frame, and in the difference code amount with respect to the target code amount, the code amount before multiplying the coefficient is expressed as follows. Needless to say, it must be used.

<< Embodiment 3 >>
Also in the third embodiment, the contents of further development of the contents described in the first embodiment will be described.

In the first embodiment, when the stop block detection unit 100 sets the position of the stop block in the encoded frame and the next moving image encoding unit 200 performs encoding, the block set as the stop block Explained that a part of the encoding process is skipped.

However, in the case of this configuration, the processing in the stop block detection unit 100 and the processing in the moving image encoding unit 200 in FIG. 1 are executed sequentially (see FIG. 10). In the third embodiment, a method for realizing higher speed will be described.

In Embodiment 1, FIG. 1 is used to show an example in which a stop block detection unit 100 that sets stop blocks and a moving image encoding unit 200 that executes encoding processing are configured by different means. It was. On the other hand, in the coding processing apparatus 300b of the present embodiment, as shown in FIG. 11, the setting of the stop block and the coding processing are executed by the same means, that is, the moving picture coding unit 200b, and s is set. When an integer is used, the stop block of the s + 1 frame is set in the motion search process executed by the motion search unit 201 when performing the encoding process of the s frame (see FIG. 12). .

Specific explanation. An input image and a reference image are input to the motion search unit 201b in FIG. 12, and a motion search is executed. At that time, the movement amount shown in Equation 1 is also calculated at the same time, and the result is stored in the movement amount storage unit 102b. When the processing for one frame (movement amount detection processing) is completed, the movement amount information stored in the movement amount storage unit 102b is sorted in ascending order of movement amount, and is output to the stop block selection unit 103b in ascending order. The stop block selection unit 103b sets stop blocks corresponding to the number of corresponding block settings, and outputs the stop block information to the moving image encoding unit 200b.

The relationship between p, m, and n when setting a stop block in the stop block selection unit 103b is not “stop block detection using a reduced image”, so that “p = m, n = 1”. Needless to say.

Thus, according to the third embodiment of the present invention, the stop block information set by the stop block selection unit 103b is used as stop block position information when processing the next frame. That is, by executing the stop block detection process and the moving picture encoding process in parallel, the process can be completed at a higher speed (see FIG. 13).

The encoding processing apparatuses 300 and 300b according to Embodiments 1 to 3 of the present invention have been described above. Next, application examples of these encoding processing apparatuses 300 and 300b will be described.

FIG. 14 is a configuration block diagram of an imaging system 601 using the encoding processing apparatus according to the present invention, for example, a digital still camera (DSC). A signal processing device 606 in FIG. 14 is one of the encoding processing devices 300 and 300b according to the first to third embodiments of the present invention.

According to FIG. 14, the image light incident through the optical system 602 forms an image on the sensor 603. The sensor 603 is driven by the timing control circuit 609 to accumulate the imaged image light and photoelectrically convert it into an electrical signal. An electrical signal read from the sensor 603 is converted into a digital signal by an analog / digital converter (ADC) 604 and then input to an image processing circuit 605 including the signal processing device 606. The image processing circuit 605 performs image processing such as Y / C processing, edge processing, image enlargement / reduction processing, and image compression / expansion processing using the present invention. The image-processed signal is recorded or transferred to a medium in a recording / transfer circuit 607. The recorded or transferred signal is reproduced by the reproduction circuit 608. The entire imaging system 601 is controlled by a system control circuit 610.

Note that the image processing in the signal processing device 606 according to the present invention is not necessarily applied only to a signal based on the image light imaged on the sensor 603 via the optical system 602, and is input as an electrical signal from an external device, for example. Needless to say, the present invention can also be applied to processing an image signal to be processed.

As described above, since the encoding processing method and the encoding processing apparatus according to the present invention can complete the encoding processing at high speed, the encoding processing method and the encoding processing apparatus are useful for high-spec imaging systems and the like having a high frame rate and high resolution. It is.

100 stop block detection unit 101 movement

amount detection unit

102, 102b movement

amount storage unit

103, 103b stop block selection unit 110 input image reduction unit 111 reference

image reduction unit

200, 200b moving

image encoding unit

201, 201b motion search unit 202 motion Compensation image generation unit 203 Subtractor 204 Selector 205 DCT processing unit 206 Quantization processing unit 207 Inverse quantization processing unit 208 Inverse DCT processing unit 209 Variable length encoding unit 211 Switch 212 Adder 213 Reference frame storage unit 300, 300b Encoding Processing device 601 Imaging system 605 Image processing circuit 606 Signal processing device (encoding processing device)

Claims

A processing target frame is divided into a plurality of blocks, a motion vector is generated for each block from an input image and a reference image that has already been encoded, and a motion vector is generated using the motion vector. An encoding processing method for performing image encoding processing,
Selecting S (S is an integer greater than or equal to 1) blocks from among the blocks in the frame;
A coding process comprising: performing the moving picture coding process on the selected S blocks without executing the motion search and forcibly replacing the motion vector with 0. Method.
An encoding process in which a processing target frame is divided into a plurality of blocks, a DCT process and a quantization process are performed for each block to generate a DCT coefficient, and a moving picture encoding process is performed using the DCT coefficient A method,
Selecting S (S is an integer greater than or equal to 1) blocks from among the blocks in the frame;
For the selected S blocks, the DCT processing and the quantization processing are not executed, and a DCT coefficient is forcibly replaced with 0 and a moving image encoding processing is performed. An encoding processing method.
The encoding processing method according to claim 1 or 2,
Detecting a moving amount of the image for all or a part of the blocks in the frame;
A coding processing method characterized by comparing movement amounts of all or some of the blocks, and selecting S blocks from the smaller movement amount.
The encoding processing method according to claim 3, wherein
Further comprising the step of reducing the frame image to 1 / n (n is an integer of 2 or more),
An encoding processing method, wherein the amount of movement of the image is detected using the reduced image.
The encoding processing method according to claim 1 or 2,
An encoding processing method, comprising: executing a moving image encoding process by forcibly replacing a DCT coefficient with 0 only r−1 times in r frames (r is an integer of 2 or more).
The encoding processing method according to claim 5, wherein
Multiplying the code amount of the previous frame by the total number of blocks / S to correct the code amount;
And a step of performing bit rate control based on the corrected code amount.
A processing target frame is divided into a plurality of blocks, a motion vector is generated for each block from an input image and a reference image that has already been encoded, and a motion vector is generated using the motion vector. An encoding processing method for performing image encoding processing,
A first step of detecting a movement amount of the image and storing it in a movement amount storage unit;
A second step of comparing the movement amounts in the movement amount storage unit and selecting S (S is an integer of 1 or more) blocks from the smaller movement amount;
For the selected S blocks, the motion search is not performed, and a third step is performed in which a motion vector is forcibly replaced with 0 and a moving image encoding process is performed.
An encoding processing method characterized in that, at the same time as executing the third step of a certain frame, the movement amount used in the processing of the third step of the next frame is obtained in the first and second steps.
An encoding process in which a processing target frame is divided into a plurality of blocks, a DCT process and a quantization process are performed for each block to generate a DCT coefficient, and a moving picture encoding process is performed using the DCT coefficient A method,
A first step of detecting a movement amount of the image and storing it in a movement amount storage unit;
A second step of comparing the movement amounts in the movement amount storage unit and selecting S (S is an integer of 1 or more) blocks from the smaller movement amount;
For the selected S blocks, the DCT processing and the quantization processing are not performed, and a third step is performed in which a DCT coefficient is forcibly replaced with 0 and a moving image encoding processing is performed.
An encoding processing method characterized in that, at the same time as executing the third step of a certain frame, the movement amount used in the processing of the third step of the next frame is obtained in the first and second steps.
The encoding processing method according to claim 7 or 8,
The movement amount storage unit can store only the S movement amount information,
When storing the result of detecting the movement amount, the movement amount is compared with the movement amount already stored in the movement amount storage unit, and if the value is smaller, it is updated and stored in the movement amount storage unit. An encoding processing method characterized by not updating and storing a large value.
A processing target frame is divided into a plurality of blocks, a motion vector is generated for each block from an input image and a reference image that has already been encoded, and a motion vector is generated using the motion vector. An encoding processing device that performs image encoding processing,
An input image reduction unit that generates an image reduced to 1 / n (n is an integer of 1 or more) with respect to an input image to be encoded;
A reference image reduction unit that generates an image reduced to 1 / n with respect to a reference image that has already been encoded;
A movement amount detection unit that inputs a reduced input image and a reference image from the input image reduction unit and the reference image reduction unit, and detects a movement amount for each block;
A movement amount storage unit for storing the movement amount detected by the movement amount detection unit;
A stop block selection unit that compares the movement amount stored in the movement amount storage unit and selects S blocks (S is an integer of 1 or more) from the smaller movement amount;
For the selected S blocks, there is provided a moving image encoding unit that does not execute the motion search and forcibly replaces a motion vector with 0 to execute a moving image encoding process. A characteristic encoding processing apparatus.
An encoding process in which a processing target frame is divided into a plurality of blocks, a DCT process and a quantization process are performed for each block to generate a DCT coefficient, and a moving picture encoding process is performed using the DCT coefficient A device,
An input image reduction unit that generates an image reduced to 1 / n (n is an integer of 1 or more) with respect to an input image to be encoded;
A reference image reduction unit that generates an image reduced to 1 / n with respect to a reference image that has already been encoded;
A movement amount detection unit that inputs a reduced input image and a reference image from the input image reduction unit and the reference image reduction unit, and detects a movement amount for each block;
A movement amount storage unit for storing the movement amount detected by the movement amount detection unit;
A stop block selection unit that compares the movement amount stored in the movement amount storage unit and selects S blocks (S is an integer of 1 or more) from the smaller movement amount;
A moving picture coding unit that executes the moving picture coding process by forcibly replacing the DCT coefficient with 0 without performing the DCT process and the quantization process on the selected S blocks; An encoding processing apparatus comprising:
The encoding processing device according to claim 10 or 11,
An encoding processing apparatus that executes a moving image encoding process in which a DCT coefficient is forcibly replaced with 0 only r-1 times in r frames (r is an integer of 2 or more).
The encoding processing device according to claim 12,
An encoding processing apparatus, further comprising means for correcting the code amount by multiplying the code amount of the previous frame by the total number of blocks / S, and performing bit rate control based on the corrected code amount.
An image processing circuit comprising the encoding processing device according to claim 10 or 11,
A sensor that converts image light into an image signal;
An optical system that forms an image of incident image light on the sensor;
An imaging system comprising: a converter that converts the image signal into digital data and outputs the digital signal to the image processing circuit.
An image processing circuit comprising the encoding processing device according to claim 10 or 11,
A signal processing system comprising: a converter that converts an input analog value image signal into digital data and outputs the digital data to the image processing circuit.